Re: Namespaces, schemas, Simon's filters.

[Joe English <jenglish@flightlab.com>]

Jeff Lowery wrote:

> (Do namespaces have names, or do names belong to
> a namespace? There's a practical representational difference but is there a
> semantic one?)

Neither: element names and attribute names have namespace names.

You can take "have" to be a symmetric relation and say the converse,
that namespaces have names -- if QName x has namespace y, then
namespace y "has" name x -- but I think this is misleading.
Given a QName x and a set of in-scope namespace declarations,
you can determine the namespace y unambiguously; but given a
namespace name (URI) y, there's no way in general to tell what
names "belong to" it.

People often say "element z is in namespace y" as shorthand
for "the namepspace name of the generic identifier x of the
element z is the URI y".  Personally I think this usage
confuses the issue, but it's a convenient abbreviation.


--Joe English

  jenglish@flightlab.com